train_multirc_1754652151

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the multirc dataset. It achieves the following results on the evaluation set:

• Loss: 0.3333
• Num Input Tokens Seen: 132272272

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 123
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.4234	0.5	3065	0.4323	6639424
0.3074	1.0	6130	0.3545	13255424
0.3571	1.5	9195	0.3608	19871232
0.4006	2.0	12260	0.3467	26471216
0.3218	2.5	15325	0.3434	33075856
0.3285	3.0	18390	0.3429	39694112
0.3234	3.5	21455	0.3399	46313216
0.3171	4.0	24520	0.3392	52929744
0.3376	4.5	27585	0.3455	59549072
0.3583	5.0	30650	0.3383	66152480
0.31	5.5	33715	0.3388	72765696
0.3494	6.0	36780	0.3389	79389648
0.3433	6.5	39845	0.3398	86008784
0.3374	7.0	42910	0.3414	92621824
0.3568	7.5	45975	0.3353	99237152
0.3462	8.0	49040	0.3341	105830544
0.3273	8.5	52105	0.3358	112458064
0.3189	9.0	55170	0.3333	119047920
0.3519	9.5	58235	0.3348	125686064
0.388	10.0	61300	0.3337	132272272

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1